OpenAI has to initially limit access to its latest AI model, GPT-5.6, to a small group of partners at the U.S. government's request.
CEO Sam Altman announced this during an internal Q&A session on Wednesday, according to a report by The Information. The U.S. federal government had "asked" OpenAI to do so. The request likely stems from the Trump administration's recently published executive order, which calls for a voluntary review of new AI models, especially around cybersecurity.
The OpenAI case now shows just how voluntary that review really is. In the memo, Altman explained that the government would approve access "on a customer by customer basis" during the preview phase. He hopes for a broader release a "couple of weeks later," assuming everything goes smoothly.
According to The Information, the push for a phased release came out of talks with two government agencies, the Office of the National Cyber Director and the Office of Science and Technology Policy.
"We’ve made clear to the U.S. government that this is not our preferred long term model, and will work with them and others in industry to achieve a more sustainable approach for future releases," Altman wrote in the memo.
Even after OpenAI had shared its plans for the limited release with senior government officials, Altman got a call from Commerce Secretary Howard Lutnick, according to The Information. Lutnick warned the company against moving forward without sign-off from more agencies.
The current situation traces back to Anthropic's "Mythos" presentation in early April. During the limited launch, the AI company put heavy emphasis on cybersecurity risks and said it would release the model only in phases. Independent groups and companies confirmed some of those risks. That got the attention of the U.S. government, which saw both opportunities and threats for its own cybersecurity.
After Anthropic released Fable, the first public model in the Mythos class, along with specific security guidelines, the U.S. government stepped in and forced the company to pull it offline. Anthropic had reportedly been working with the government beforehand to find security flaws in Fable, and according to The Information, authorities raised no objections when the company announced the planned release.
Things still blew up after launch. One likely factor is the tense relationship between Anthropic and the Pentagon. The company was classified as a supply chain risk after it refused to give the Pentagon access for domestic mass surveillance and autonomous warfare.
Talks about re-releasing Fable are ongoing. A White House spokesperson told The Information, "The White House continues to collaborate with frontier AI labs to develop shared approaches for addressing the challenges of scaling this technology."
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。